Escolar Documentos
Profissional Documentos
Cultura Documentos
Design Outcomes
This was a naturalistic, prospective study. Measures Outcome measure 1 referred to academic perform-
were taken at the beginning of the psychiatry ance, measured by students’ scores (%) in the 2
attachment (baseline) and at the end of the attach- routine formative examinations conducted on the
ment (outcome). The 2 cohorts of medical students last day of the attachment: an MCQ test and a clinical
(PBL and traditional curricula) were compared. viva. The pass mark was 50% for both.
Baseline measure 3 concerned student learning Outcome measure 4 referred to social desirability,
styles, measured using the 28-item version of the measured using the 13-item version of the Marlowe-
Study Process Questionnaire23 (SPQ). This consists of Crowne scale,25 which is valid and reliable when used
3 subscales, which indicate the extent to which with undergraduate students. This scale measures
students use surface, deep, or strategic learning the tendency to give answers that are perceived as
styles. The SPQ is reliable and acceptable when used acceptable to the investigator rather than
with medical students.17,24 representing the respondent’s true opinion.
Table 1 Mean (SD) scores for examination performance, attitudes to psychiatry and learning styles
Cohort 1 Cohort 2 Total
Measure (n ¼ 188) (n ¼ 191) (n ¼ 379)
Examination performance
MCQ (% score) 60.7 (9.7) 64.3 (10.9) 62.5 (10.5)
Viva (% score) 69.8 (15.6) 73.6 (15.5) 71.7 (15.7)
ATP-30
Baseline 102.7 (10.6) 102.6 (9.6) 102.7 (10.1)
Outcome 108.4 (11.9) 107.0 (11.8) 107.7 (12.0)
SPQ surface learning
Baseline 14.1 (3.7) 14.5 (3.8) 14.3 (3.8)
Outcome 14.0 (3.8) 14.6 (3.8) 14.2 (3.8)
SPQ deep learning
Baseline 19.4 (4.3) 18.8 (3.6) 19.1 (4.0)
Outcome 19.4 (4.6) 18.7 (3.9) 19.1(4.3)
SPQ strategic learning
Baseline 16.1 (4.7) 16.3 (4.5) 16.2 (4.6)
Outcome 16.4 (4.7) 16.2 (4.3) 16.3 (4.5)
split-plot ANOVA, with 1 within-subjects factor (time of the end of the attachment for each of the subscales
administration of ATP-30, which had 2 levels, base- were predicted by their scores at the beginning of the
line and outcome) and 1 between-subjects factor attachment (P < 0.001 for all 3 subscales). The
(student cohort, which had 2 levels, Cohort 1 and amount of variance in the outcome scores accounted
Cohort 2). At the end of the attachment students’ for by the baseline scores (adjusted R squared) was
attitudes were found to have improved over the found to be 41% for surface learning, 57% for deep
course of the attachment for both cohorts (F ¼ 91.1, learning, and 66% for strategic learning.
P < 0.001). The mean change in ATP-30 total score
was 5.0 (SD ¼ 10.2). The 2 cohorts did not differ Very low correlations were found between the surface
from each other at either the beginning or end of the and deep subscales (r ¼ ) 0.13 at baseline and
attachment and there was no interaction between r ¼ )0.10 at outcome), and between the surface and
cohort and change over time. strategic subscales (r ¼ 0.09 at baseline and r ¼ 0.17
at outcome). However, the deep and strategic sub-
Effects of type of curriculum on learning styles scales correlated modestly at both times of adminis-
tration (r ¼ 0.36 at baseline and r ¼ 0.40 at outcome,
The alpha reliabilities for the 3 subscales of the both P < 0.001). This indicates that although the
18-item SPQ were comparable to those found by subscales were relatively independent, there was an
McManus et al.15 The reliabilities (baseline and out- association between the use of deep learning and the
come) for surface learning were 0.59 and 0.63, for use of strategic learning.
deep learning were 0.69 and 0.77, and for strategic
learning were 0.67 and 0.70. These indicated that the To compare surface, deep and strategic learning
subscales had modest to good reliability. styles by time of administration and cohort (Table 1),
split-plot ANOVA was performed with 1 within-subjects
Correlations were found between scores on the factor (time of administration, with 2 levels, baseline
subscales at the beginning and the end of the and outcome) and 1 between-subjects factor (student
attachment. Correlations were high for deep learning cohort, with 2 levels, Cohort 1 and Cohort 2). No
(r ¼ 0.76, n ¼ 379, P < 0.001) and for strategic differences were found in surface, deep or strategic
learning (r ¼ 0.81, n ¼ 379, P < 0.001), and modest learning between students at the beginning and end
for surface learning (r ¼ 0.64, n ¼ 379, P < 0.001). of the attachment. No differences were found in
Regression analysis indicated that students’ scores at surface, deep or strategic learning between the 2
Table 2 Correlations between SPQ subscales and examination scores for the total sample (n ¼ 379)
Examination Correlation Surface Deep Strategic
MCQ Pearson’s r 0.09 0.03 0.14*
Viva Spearman’s rho ) 0.07 0.11 0.11
* P < 0.01; P < 0.05.
cohorts. There were no interactions between time of 4 The type of curriculum did not affect students’
administration and cohort for surface, deep or attitudes to psychiatry or their learning styles at
strategic learning. the end of the attachment.
5 Students’ attitudes to psychiatry at the end of the
Correlations between the 3 learning styles at the end attachment were predicted by their initial atti-
of the attachment and performance in the MCQ and tudes (prior to any psychiatry teaching) and their
viva examinations were all found to be very low learning styles at the end of the attachment were
(Table 2). Out of the 3 learning styles, only the predicted by their initial learning styles.
strategic subscale correlated significantly with both 6 Student characteristics (age and gender) showed
measures of academic performance. some relationships to students’ initial attitudes
and learning styles. Female students had more
Predictors of examination performance, attitudes to favourable attitudes to psychiatry at the outset and
psychiatry and learning styles were more likely to use a strategic learning style.
Older students were less likely to use a surface
Multiple linear regression analysis was conducted to learning style.
identify the relationships between students’ initial
and end-of-attachment attitudes and learning styles,
examination performance and student characteris- DISCUSSION
tics. This analysis (a form of structural equation
modelling) enables a diagram to be constructed The results indicate that the PBL curriculum was more
indicating the presence and strength of relation- effective in helping students to learn, and that this
ships among the variables, and can be used to applied to both clinical performance and knowledge
examine possible causal links. The results are shown base. This is encouraging, as it shows that the acqui-
in Fig. 1. sition of knowledge need not be a casualty of curricu-
lum reform. Although students were not randomised
In summary, the analysis indicates that: to the 2 cohorts, they were equivalent with regard to
gender mix, age, attitudes to psychiatry and learning
1 Success in the MCQ was predicted by the type of styles prior to the attachment. Furthermore, both
curriculum studied (PBL students achieved high- cohorts of students were selected into medical school
er marks) (P < 0.001) and the use of a strategic in the same way, as the full-scale new curriculum had
learning style (P < 0.01). These variables accoun- not been introduced, and psychiatry was the first
ted for 5% of the variance (adjusted R squared). clinical subject to change its method of teaching.
2 Success in the clinical viva was predicted by the There is no reason, therefore, to assume that the
type of curriculum studied (PBL students difference in academic performance was due to
achieved higher marks) (P < 0.05), the use of a differences between the cohorts. For the purposes of
strategic learning style (P < 0.05) and gender ecological validity, the existing assessment procedures
(being female) (P < 0.01). These variables were used to compare the 2 cohorts. This method has
accounted for 5% of the variance (adjusted R advantages and disadvantages, as both methods of
squared). Female students achieved a mean score assessment have strengths and weaknesses. The MCQ
of 74.5% (SD ¼ 14.2) and male students achieved paper was administered blind as it was computer-
a mean score of 69.5% (SD ¼ 16.5) in the viva. marked and consisted of a large number of questions
3 Students’ attitudes to psychiatry did not predict covering the entire syllabus. Furthermore, students in
their performance in either the MCQ or the viva each block were given different papers, so they could
examinations. not learn the questions from their peers earlier in the
0.58
Attitudes 1 Attitudes 2
Cohort 0.18
Age -0
.16
MCQ
(PBL)
4
0.1
Surface 0.64 Surface
7
learning 1 learning 2
0.1
-0.08 0.
12
Key
Lines represent predictive
relationships
= positive relationship
= negative relationship
No line = no relationship
Numbers are beta weights
Figure 1 Diagram constructed using multiple linear regression analysis to indicate the relationships between
students’ initial and end-of-attachment attitudes and learning styles, examination performance and student
characteristics.
year. As the same set of rotating papers was used in the There were no improvements in attitudes to psychi-
2 cohorts, the standard of the papers used was the same atry in the PBL cohort compared to the traditional
for both years. However, the use of different papers in curriculum cohort, which corroborates findings from
each cohort, providing the same standard was guar- another study on a psychiatry attachment,9 although
anteed, would have ruled out any passing of questions it is not consistent with the general PBL literature.1–3
between the 2 cohorts. The clinical viva was conducted Attitudes in both cohorts improved during the course
with clear guidelines and marking criteria by pairs of of the attachment. This may be because change in
experienced examiners, and although some of the attitudes is driven more by students’ clinical experi-
examiners were not blind to which teaching method ences (such as encouragement from consultants and
had been used, minutes from the teachers’ meetings seeing patients respond well to treatment) than by
indicate that they did not expect students in the PBL formal teaching sessions.8
cohort to do better in the examinations. However, vivas
are known to be somewhat unreliable in assessing Contrary to expectations, improved academic per-
knowledge, and a standardised procedure with blind formance was found without any changes in learning
marking would have helped to corroborate the find- styles, and was unrelated to students’ attitudes.
ings. In addition, the expectations of examiners about Consistent with previous research, academic success
the performance of the 2 cohorts were not formally was related to the use of strategic and deep learning
assessed and compared to the marks given, which styles. It is possible that PBL improved students’
might have ruled out potential effects of examiner ability to learn during the teaching sessions, but did
expectation bias. not lead to a change in students’ preferred approach
to learning outside the sessions. The findings there-
Success in the clinical viva was related to being fore suggest that exposure to PBL methods in a single
female. This is consistent with other findings in the 8-week attachment is not sufficient to alter students’
literature ascertaining that women tend to perform preferred learning styles, but that increases in the
better than men in their medical training, partic- effectiveness of learning can still be achieved. The
ularly in clinical assessments.25 lack of association between attitudes to psychiatry and
success in psychiatry examinations may also be a 5 Colliver JA. Effectiveness of problem-based learning
consequence of students’ having a preferred way of curricula: research and theory. Acad Med 2000;75:
approaching examinations. It may be that because 259–66.
students have to take examinations in certain sub- 6 Smits PBA, Verbeek JHAM, de Buisonje¢ CD.
Problem-based learning in continuing medical
jects, whether they like the subjects or not, they
education: a review of controlled evaluation studies.
develop an approach to revising which they use
BMJ 2002;324:153–6.
irrespective of the style of teaching or their degree of 7 Kaufman DM, Mann KV. Comparing achievement
interest in the subject. on the Medical Council of Canada Qualifying
Examination Part 1 of students in conventional
In summary, the findings indicate that the PBL course and problem-based learning curricula. Acad Med
was more successful in terms of students’ academic 1998;73:1211–3.
performance than the traditional course, and this 8 McParland M, Noble LM, Livingston G, McManus C.
suggests that the change to a PBL course is worthwhile. The effect of a psychiatric attachment on students’
attitudes to and intention to pursue psychiatry as a
career. Med Educ 2003;37:447–54.
9 Singh SP, Baxter H, Standen P, Duggan C. Changing
CONTRIBUTORS
the attitudes of ÔtomorrowÕs doctors’ towards mental
illness and psychiatry: a comparison of two teaching
GL, MM and LN designed the study and submitted methods. Med Educ 1998;32:115–20.
the protocol for ethical approval. MM collected and 10 Wilkinson DG, Greer S, Toone BK. Medical
inputted the data, under the supervision of GL and students’ attitudes to psychiatry. Psychol Med 1983;13:
LN. All 3 authors were involved in the data analysis 185–92.
and the write-up of the manuscript. 11 Alexander DA, Eagles JM. Changes in attitudes towards
psychiatry among medical students: correlation of
attitude shift with academic performance. Med Educ
ACKNOWLEDGEMENT 1990;24:452–60.
12 Marton F, Säljö R. On qualitative differences in learn-
ing. I. Outcome and process. Br J Educ Psychol
The authors wish to thank the medical students at
1976;46:4–11.
UCL for participating in this research.
13 Newble DI, Entwistle NJ. Learning styles and approa-
ches: implications for medical education. Med Educ
1986;20:162–75.
FUNDING 14 Fransson A. On qualitative differences in learning. IV.
Effects of intrinsic motivation and extrinsic test anxiety
None. on process and outcome. Br J Educ Psychol 1977;47:244–
57.
15 McManus IC, Richards P, Winder BC, Sproston KA.
ETHICAL APPROVAL Clinical experience, performance in final examina-
tions, and learning style in medical students: pros-
pective study. BMJ 1998;316:345–50.
This study was approved by the Joint UCL ⁄ UCLH
16 Coles CR. Differences between conventional and
Ethical Committee.
problem-based curricula in their students’ approaches
to studying. Med Educ 1985;19:308–9.
17 Newble DI, Clarke RM. The approaches to learning of
REFERENCES students in a traditional and in an innovative problem-
based medical school. Med Educ 1986;20:267–73.
1 Albanese MA, Mitchell S. Problem-based learning: a 18 Mennin SP, Kalishman S, Friedman M, Pathak D,
review of literature on its outcomes and implementa- Snyder J. A survey of graduates from the University
tion issues. Acad Med 1993;68:52–81. of New Mexico’s conventional and community-
2 Vernon DTA, Blake RL. Does problem-based learning orientated problem-based track. Acad Med 1996;71:
work? A meta-analysis of evaluative research. Acad Med 1079–89.
1993;68:550–65. 19 Newble DI, Jaeger K. The effect of assessments and
3 Norman GR, Schmidt HG. The psychological basis of examinations on the learning of medical students. Med
problem-based learning: a review of the evidence. Acad Educ 1983;17:165–71.
Med 1992;67:557–65. 20 Burra P, Kalin R, Leichner P, Waldron JJ, Handforth
4 Albanese M. Problem-based learning: why curricula are JR, Jarrett FJ, Amara B. The ATP-30 – a scale for
likely to show little effect on knowledge and clinical measuring medical students’ attitudes to psychiatry.
skills. Med Educ 2000;34:729–38. Med Educ 1982;16:31–8.
21 Biggs JB. Student Approaches to Learning. Victoria, 24 Cohen L, Holliday M. Statistics for Social Scientists.
Australia: Australian Council for Educational Research London: Harper & Row 1982.
1987. 25 Ferguson E, James D, Madeley M. Factors associated
22 Fox RA, McManus IC, Winder BC. The shortened study with success in medical school: systematic review of the
process questionnaire: an investigation of its structure literature. BMJ 2002;324:952–7.
and longitudinal stability using confirmatory factor
analysis. Br J Educ Psychol 2001;71:511–30. Received 29 January 2003; editorial comments to authors 6
23 Reynolds WM. Development of reliable and valid short May 2003; accepted for publication 4 July 2003
forms of the Marlowe-Crowne social desirability scale.
J Clin Psychol 1982;138:119–25.